On Monday 27 April 2009 15:46:25 Wolfgang Denk wrote:
> In message Mike Frysinger wrote:
> > > > - * We need a wrapper for gunzip() because the parameters are
> > > > + * We need a wrapper for zunzip() because the parameters are
> > > > * incompatible with the lzo decompressor.
> > > > */
> > > > static int gzip_decompress(const unsigned char *in, size_t in_len,
> > > > unsigned char *out, size_t *out_len)
> > > > {
> > > > unsigned long len = in_len;
> > > > - return gunzip(out, *out_len, (unsigned char *)in, &len);
> > > > + return zunzip(out, *out_len, (unsigned char *)in, &len, 0, 0);
> > > > }
> > >
> > > If the only purpose of zunzip() is to be used here, then why do we not
> > > make the parameters fit the intended purpose, thus avoiding an
> > > additional wrapper?
> >
> > not sure i follow ... what do you propose changing ?
>
> The comment above seems to indicate that gzip_decompress() is only
> (?) needed because the zunzip() parameters are incompatible with
> those of the lzo decompressor.
>
> So if we create zunzip() just for this purpose, then why not create
> it in a way that the parameters are compatible with those of the lzo
> decompressor, so we can omit this whole gzip_decompress() wrapper
> function?that should be a follow up change, but the wrapper already exists today regardless of Ricardo's fixes. i dont think his changes should be held up to address that. that direction should probably cover: - fix gunzip code to use size_t's instead of unsigned long - re-order lzo arguments to match zlib rather than other way around since lzo has very few consumers -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ U-Boot mailing list [email protected] http://lists.denx.de/mailman/listinfo/u-boot
